Amendment history:Correction (November 1961)

GLUCAGON ANTIBODIES AND ANIMMUNOASSAY FOR GLUCAGON

Roger H. Unger, … , M. S. McCall, Leonard L. Madison

J Clin Invest. 1961;40(7):1280-1289. https://doi.org/10.1172/JCI104357.

Research Article

Find the latest version:

http://jci.me/104357-pdf

GLUCAGONANTIBODIES ANDAN IMMUNOASSAYFOR GLUCAGON*

By ROGERH. UNGER, ANNAM. EISENTRAUT, M. S. McCALL AND LEONARDL.MADISONWITH THE TECHNICAL ASSISTANCE OF KATHRYNR. SIMS,

LUCILLE TIMM AND LORETTAPATMAN

(From the Department of Medicine and Radioisotope Laboratory, Veterans Hospital; andthe Department of Medicine, University of Texas Southwestern Medical School,

Dallas, Texas)

(Submitted for publication February 6, 1961; accepted March 30, 1961)

Despite extensive knowledge of the amino acidcomposition of beef-pork glucagon, and of thephysiologic changes induced by its administrationto animals and man, the role of glucagon in thehomeostatic regulation of blood glucose level innormal and abnormal states remains to be defined.This fundamental gap in knowledge is the resultof the lack of a method sufficiently specific andsensitive to measure endogenous glucagon inbody fluids.

Previous efforts to assay glucagon in plasmahave been based upon ability to produce either hy-perglycemia in a living animal (1-3), glycogenoly-sis in liver slices (4), or phosphorylase activationin liver homogenates (5). There is reason todoubt the specificity of the plasma glucagon-likeactivity detectable by such methods (6), and theirstatus is at present in question.

The greater specificity inherent in immunoas-says (7), coupled with the unprecedented sensi-tivity provided by the paper radiochromato-elec-trophoretic techniques of Berson and co-workers(8) suggested that a radio-immunoassay for glu-cagon might provide a hitherto unequaled rangeof specific measurements. In a preliminary re-port, the antigenicity of beef-pork glucagon inrabbits was recorded and the feasibility of a radio-immunoassay for glucagon demonstrated (9).The development by Berson and Yalow of a ra-dio-immunoassay for insulin of remarkable sensi-tivity and specificity (10, 11) has validated thisapproach to problems of peptide hormone assay.

In the following report the antigenicity of glu-cagon is demonstrated, glucagon antibodies are

* Supported by United States Public Health ServiceGrant A-2700. Portions of this study were presentedat the Central Society for Clinical Research, Chicago,Ill., November 6, 1959, and the American Society forClinical Investigation, Atlantic City, N. J., May 2, 1960.

characterized in respect to their specificity and thekinetics of their reaction with glucagon, and aradio-immunoassay for glucagon capable of meas-uring as little as 50-millionths of a microgram ofglucagon is described.

METHODSAND MATERIALS

I. Antibody studies

A. Immunization. Forty-seven white female rabbits,each weighing more than 5 pounds, were employed inthis study. Twenty-nine were given monthly inj ectionsof Lilly beef-pork glucagon 1 in the complete Freund'sadjuvant for 3 months and at irregular and less frequentintervals thereafter; 18 control animals received simi-larly scheduled injections of either Lilly beef-pork "glu-cagon-free" insulin in complete Freund's adjuvant (7rabbits) or complete Freund's adjuvant alone (5 rab-bits), and 6 rabbits were given no injections at all.

Premixed Lilly beef-pork glucagon solution was addedwith sterile technique to an equal volume of the com-plete Freund's adjuvant and mixed at room temperatureon an automatic paint shaker. At least 15 minutes wasrequired for emulsification; 2 ml of emulsion containing1 mg of glucagon was administered to each rabbit; 0.2 mlwas injected in a footpad and the remainder subcutane-ously in a hindquarter.

B. Radioiodination of glucagon. The procedure em-ployed is a modification of the Pressman-Eisen technique(12) and is being reported elsewhere in greater detail(13). To 50 me of Nal"', chilled in a beaker of ice, 3drops of 0.1 NaNO, are added; 2.5 N HCl is added drop-wise, to give a pH of about 2.0. Then 0.05 ml of 0.001N KI and 1 drop of NaNO, are added; 50 jug (0.1 ml)of chilled glucagon is rapidly added. Prompt addition of2 to 3 ml of glycine buffer with thorough but gentle mix-ing follows, and the pH is raised to 9.0 or above. Thematerial is then dialyzed for 3 hours in 250 ml of chilledglycine buffer containing 15 ml of washed resin (Amber-lite 401). The buffer is maintained in an ice bath andthe resin gently agitated by a magnetic stirrer. Thedialyzed material is transferred to a tube containing

1 Kindly supplied by Dr. W. R. Kirtley, Eli Lilly, Inc.,Indianapolis, Ind.

1280

GLUCAGONANTIBODIES AND AN IMMUNOASSAYFOR GLUCAGON

enough albumin to provide a final albumin concentrationof 5 per cent. This solution is passed through a Resikitto reduce further the nonprotein-bound radioactivity.The final glucagon-1l" solution is frozen immediately andstored until needed.

The technique described provides a high specific ac-

tivity without undue damage to the glucagon molecule.Biologic assays of one lot of glucagon-Il31, kindly per-formed by Dr. W. Bromer of Eli Lilly, Inc., failed todemonstrate loss of biologic activity despite many weeksof storage. In general, an atom: molecule ratio of 0.7: 1can be anticipated.

C. Detection of antibodies. Serum obtained from eachrabbit prior to the monthly injection was studied by theBerson, Yalow and Volk technique adapted for glucagon-IP31 (14). This method is based upon the observationthat glucagon-I1", when incubated in normal serum andthen applied to a filter-paper strip, is adsorbed to thepaper at the point of its application (origin) and will not,under these circumstances, migrate either chromato-graphically or electrophoretically, thereby permitting itsseparation from the serum proteins. In the presence ofglucagon antibodies, however, glucagon-I13. might be ex-

pected to bind to and migrate with globulin in a mannner

similar to that described for insulin-I.31 and insulin anti-bodies (8).

One ml of undiluted serum was incubated on a rotatingshaker at 370 C for 1 to 3 hours with 10 m/ug of glucagon-I". In the more recent studies this has been followedby overnight incubation at 40 C. From 20 to 100 Al ofthis mixture was then applied to a Whatman 3 MMfilter-paper strip. In a barbital buffer, ionic strength 0.05,pH 8.6, the strips were subjected to electrophoresis withhydrodynamic flow 2 in a Spinco model R electrophoresiscell with a current of 10 ma for 1 hour, followed by elec-trophoresis alone at 2.5 ma for 16 hours. Strips were

heat-dried at 1200 C for 1 hour and stained with naphtholblue-black dye. The strips were counted in an Actigraphcounter with a TGC-2 Geiger-Muller tube. Count-ratetracings were obtained by means of a Rectiriter recorder(Texas Instruments).

In other studies not requiring separation of the indi-vidual serum protein fractions, the strips were loadedwith up to 300 ,l and subj ected to horizontal chroma-tography for periods up to 2.5 hours, during which timethe proteins migrate at least 2 inches from the origin.After drying and staining, the strips were cut so as toseparate completely the origin from the protein area.

Each half of the strip was counted in a well-type scintil-lation counter. All samples were counted until a totalof at least 10,000 counts was obtained. Berson, Yalow,and Volk have shown that a small fraction of glucagon-I... damaged by irradiation or other factors, during or

after the iodination procedure, migrates nonspecificallywith serum proteins (14). Parallel control runs em-

2 Hydrodynamic flow occurs if the slit at the top ofthe cell cover is left open. The resulting evaporationprovides the necessary flow to move y-globulin away

from the origin.

ploying undiluted serum from unimmunized rabbits were,therefore, carried out with every run in order to distin-guish between nonspecific protein-binding of damagedglucagon-I13' and specific antibody-binding of undamagedglucagon-I". The damaged moiety ranged from 4.8 to20 per cent of the glucagon-1l"' lots employed in thisgroup of studies, but was almost always less than 10per cent.

In addition to the Berson-Yalow techniques, the Skom-Talmage technique (15), originally introduced for thestudy of nonprecipitating insulin antibodies, was adaptedwithout change except that glucagon-I31, glucagon-chal-lenged rabbit serum, and sheep antirabbit globulin serumwere substituted for insulin-I"3', human anti-insulin serum,and antihuman globulin serum, respectively.

II. Immunoassay for glucagonA. Principles of the radio-immnunoassay. The reac-

tion between radioantigen and antibody can be quanti-tatively expressed as a ratio of the migrating or anti-body-bound glucagon-I.1. to the nonmigrating or free glu-cagon-113l. This ratio will be referred to henceforth asthe B/F ratio. It was shown previously that, if theamount of glucagon-I13' and antibody in the system iskept constant, the introduction of progressively increas-ing amounts of nonradioactive glucagon, by competingwith glucagon-I.3. for antibody, will cause a progressivedecrease in the B/F ratio of glucagon-I.1. (9). As inthe Yalow-Berson (11) and Grodsky-Forsham (16) in-sulin assays, it is the relationship between the quantityof unlabeled antigen introduced and the B/F ratio of theradioantigen which forms the basis of the immunoassay.

B. Materials. Antiserum. All rabbit sera capableof binding 50 per cent or more of 10 m~ug of glucagon-I13.per ml were pooled, to assure a large stockpile of serumof uniform antibody concentration.

Glucagon standards. Solutions of unlabeled beef-porkglucagon premixed in Lilly glucagon diluent 3 were madeup in 5 per cent human albumin solution in concentrationsof from 50 to 10,000 jAg per ml. These can be storedfrozen without notable loss of glucagon.

Glucagon-1111. Batches of glucagon-I'3' with specificactivities ranging from 93 to 234 mc per mg were pre-pared as described by the McCall modification (13) ofthe Pressman-Eisen (12) technique. Five per cent albu-min solution was added to the glucagon-113. to give astock solution of 1 or 2 ,g per ml. The albumin pre-vents adsorption to glassware and minimizes the radia-tion damage (14) which ranged from 2.06 to 9.60 percent in the immunoassay studies. Frozen glucagon-I"3can be stored without significant increase in damage forup to 2 weeks. A trace concentration of 50 to 100,ug of glucagon-I131 per 0.1 ml was obtained by diluting25 to 100Al of stock solution with 2 per cent albuminsolution.

3 Formula of the diluent, kindly supplied by Dr. W. R.Kirtley, is as follows: 981.73 g of distilled water, 16 gof glycerin, and 2.27 g of phenol. The pH is adjusted to2.85 with HCl.

1281

R. H. UNGER, A. M. EISENTRAUT, M. S. McCALL AND L. L. MADISON

- 10.5

7. 5

4 5

1 5

CONTRORABBIT SEI

Or g i n

.:..: .:A

c

z

-0

en

z

ce

3,0

GLUCAGONClH

1.8/ \ RABBIT SE

+

Origin ALE

C e n t me t e r 5

FIG. 1. CONTROL AND GLUCAGON-CHALLENSERA. A. When glucagon-I13. is incubatedfrom normal control rabbits, and then subjectrophoresis plus hydrodynamic flow, the ma

of radioactivity remains fixed to the origin (

giving the above pattern of distribution of r

B. With serum of rabbits challenged witglucagon, the pattern of distribution of i

changes strikingly with a substantial fracti(grating with the -y-globulins.

Unknown specimens. In these studies unkmens consisted either of heparinized humancrude pancreatic extracts. All blood was

immediately after collection, and the plasmarated and promptly frozen until the timeExtracts of pancreatic tissue were made byof Sutherland and de Duve (17), and the

residue reconstituted in chloride-phosphate buffer con-taining 5 per cent albumin.

L C. Procedure. 1) One ml of a 1: 100 dilution of pooledRUM high titer rabbit antiserum in physiologic saline was in-

cubated with 1 ml of each glucagon standard (or of anunknown specimen) for 1 hour at 370 C in an automaticshaker. 2) One-tenth ml of glucagon-I13. solution, con-taining 50 to 100 Iqgg of glucagon-I.3. was added to eachtube in identical quantities and reincubated with shaking

L B for another hour at 370 C. 3) The mixtures were thenincubated at 4° C for at least 24 hours. 4) After incuba-tion the mixtures were allowed to reach room tempera-ture, and 0.9 ml of each mixture was loaded on threeWhatman 3 MMfilter-paper strips, 0.3 ml per strip, ar-ranged in horizontal chromatography cells containing abarbital buffer at pH 8.6, ionic strength 0.05. 5) Thesewere chromatographed at room temperature for ap-proximately 2.5 hours or until the proteins had migratedto the far end of the strip, their posterior edges at least2 inches beyond the anterior edge of the origin. A dummystrip containing bromphenol blue dye was used to deter-mine when migration was complete. 6) The strips wereheat-dried for 1 hour at 120- C and stained withnaphthol blue-black dye and air-dried. 7) They werethen cut so as to separate clearly the origin from the pro-tein area at the opposite extreme of the strip. The threehalves containing the origin and the three halves con-taining the protein area were carefully folded togetherand counted in a plastic tube in a 4 cc well-type scintil-

A LL E NGE D lation counter until a total of at least 10,000 counts was

obtained.Several of the glucagon standards and unknown sam-

ples were set up in the identical manner except that se-rum from non-immune rabbits was substituted for anti-

B glucagon serum. These control strips were necessary todetermine the degree of nonspecific protein-bound mi-gration of glucagon-Il1 damaged during incubation with

-------- ----- the various specimens being assayed, in order to permit5 calculation of a B/F ratio corrected for damage.

D. Calculation of B/F ratio corrected for damagedTGED RABBIT glucagon-I'31. This was done in the manner described bywith serum Yalow and Berson for insulin-I" (11).

:ted to elec-jor fraction RESULTS

of the stripi . Antigen-antibody reaction

th beef-pork Antibody response. When glucagon-I'3' wasradioactivity incubated with the serum of normal control rab-

bits and the mixture applied to a filter-paper stripand subjected to either electrophoresis with hy-drodynamic flow or chromatography, all but a

:nown speci- small fraction of the radioactivity remained fixedl plasma or to the origin or point of application, and the pat-

centrifuged tern illustrated in Figure 1A was obtained.a was sepa- In the rabbits challenged with monthly injec-of the test.the method tions of beef-pork glucagon in Freund's adjuvant,

e desiccated however, a progressively increasing deviation from

1282

GLUCAGONANTIBODIES AND AN IMMUNOASSAYFOR GLUCAGON

the above pattern was observed. After the thirdor fourth injection, a varying but usually sub-stantial portion of the glucagon-I131 migratedwith y-globulin, giving the radio-electrophoreticpattern typified by Figure 1B.

In Figure 2, glucagon-I131 migration in thesera from the 29 glucagon-challenged rabbits iscompared with that of sera from 7 insulin-chal-lenged rabbits, 5 rabbits treated with Freund'sadjuvant only, and 6 untreated rabbits. Signifi-cant migration of glucagon-I131 with y-globulinwas encountered only in sera obtained from glu-cagon-challenged rabbits, averaging 58.5 per centof 10 m~ug of glucagon-I131 with a range of 13.4 to84.6 per cent. In serum from the three controlgroups an average of only 8 per cent of the glu-cagon-I131 migrated with globulin, with a rangeof 4.8 to 13.4 per cent. This small percentage ofradioactivity which migrated diffusely with serumproteins of the control rabbits represents damagedglucagon-'13' (14). These results indicate thatsignificant levels of glucagon-binding globulinsoccur only after repeated injections of glucagon,thereby indicating its antigenicity.

The quantitative authenticity of the above re-sults was verified by a different method, the Skom-Talmage technique (15), in which the radioanti-gen-antibody complex is separated from free ra-dioantigen by precipitation of the former withspecific antiglobulin serum. In the serum ofRabbit G-23, for example, in which 85 per centof the glucagon-I131 was found by the Berson-Yalow technique to be bound to globulin, 87 percent of the glucagon-I131 was found by the Skom-Talmage technique to be bound to globulin.

Cross reaction between insulin and glucagon.In order to exclude cross reactivity between insu-lin and glucagon, glucagon-I131 was incubated with

TABLE I

Tests of cross reactivity between glucagon-I1' andinsulin-I'3l and their antibodies

% Bound toRadioantigen Serum globulin

Antiglucagon 73.1Glucagon-I'3l Anti-insulin 19.8

Non-immune 17.7

Antiglucagon 19.7Insulin1-I13' Anti-insulin 90.3

Non-immune 20.7

90-

80-

-0770-

0m

60-

0Im

IV 50-0

0

1 40-

10-

Freund'sInsulin

Nothing

FIG. 2. BINDING OF GLUCAGON-I13' IN RABBIT SERUM.

The percentage of 10 mug of glucagon-I"3' which mi-grated chromatographically with globulin in the serum

of 29 rabbits challenged with monthly injections ofbeef-pork glucagon in Freund's adjuvant, is indicated on

the right (range of migration 13.4 to 84.6 per cent; mean

58.5). The per cent migration in the serum of 18 con-

trol rabbits is recorded on the left. In 6 uninjected con-

trols, migration was 4.8 to 7.2 per cent (mean 5.8) ; in7 rabbits receiving "glucagon-free" insulin in Freund'sadjuvant, migration was 7.3 to 13.4 per cent (mean 10.4);in rabbits injected with Freund's adjuvant alone migra-tion ranged from 5.9 to 12.0 per cent (mean 7.7). Theresults indicate the antigenicity of beef-pork glucagonin rabbits.

anti-insulin rabbit serum and insulin-I131 withantiglucagon serum, and binding of the radio-antigens to globulin measured chromatographi-cally. As indicated in Table I, no evidence ofcross reactivity was noted, the binding of insulin-

i13l to globulins in antiglucagon serum and ofglucagon-I131 in anti-insulin serum being no

greater than in non-immune sera.

Protection of glucagon fromn proteolytic degra-dation. It has been shown that the binding ofinsulin to insulin antibody protects the formerfrom degradation by insulinase (18). In order to

Control Rabbits Glucagon ChallengedRabbits

mean

:I0

0

000~~~~~~~~

mean-

,*I= ._ __: _-_

:-= .-0-.F

Ii1

vr

1283

R. H. UNGER, A. M. EISENTRAUT, M. S. McCALL AND L. L. MADISON

100

! 0

"60 k,-\ ss

mpg)20 WA ib w ab

bitCaGntigu oA sr anSAIquOts. were reoved at.

various times up to 3 hours thereafter and the per centmigration at these times determined. The open circlesindicate the rate of formation of the radioantigen-anti-body complex. Subaliquots of the incubating mixtureof glucagon-1131 and antiserum were removed at S dif-ferent points and swamping quantities of nonradioactiveglucagon added in order to study dissociation of theradioantigen-antibody complex. The solid circles showthe per cent bound to antibody at various times afterthe addition of the "cold" glucagon and indicate the rateof dissociation of the radioantigen-antibody complex as

antibody-bound glucagon-I... is exchanged for and di-luted out by the nonradioactive glucagon. The resultsindicate the rate of dissociation of the complex and dem-onstrate the reversibility of the antigen-antibody reaction.

determine whether glucagon antibodies similarlyprotect glucagon from degradation by plasmin,shown by Mirsky, Perisutti and Davis to degradeglucagon (19), 1 m.,gof glucagon- 11 was incu-bated in the described manner with both antiglu-cagon serum and a non-immune serum. Afterthe usual overnight incubation 1,000 U of buff eredstreptokinase (Varidase) in fresh human plasmawas added and incubated with the mixture for 90minutes at 370 C by the Mirsky method (19).In the presence of antiglucagon serum, only 8.2per cent of the glucagon-I 3n was degraded, asindicated by loss of trichioroacetic acid-precipitableradioactivity, as compared with 37.3. per cent deg-radation in the control sample containing non-im-mune rabbit serum.

Kinetic studies. The methods used by Bersonand Yalow (20), to characterize the reaction be-tween insulin and its antibody were employed inthese studies to examine the patterns of formation

and dissociation of the glucagon-antibody com-plex. Glucagon-I131 (12.3 mug) was incubatedin a 1: 10 dilution of pooled rabbit antiglucagonserum at 370 C, and aliquots were removed atvarying intervals to determine by paper chroma-tography the amount of glucagon-I131 bound toantibody after various time intervals. The re-sults of such an experiment are recorded in Fig-ure 3, and the rate of association at that tempera-ture is indicated.

To study dissociation of the glucagon-I131 anti-body complex, saturating quantities (10 jug) ofnonradioactive glucagon were added at varyingtimes to aliquots of the incubating mixture. Sam-ples of these were removed at varying intervalsand the changes in the percentage of glucagon-I131bound to antibody were determined. Reversibilityof the reaction would be indicated by a decreasein antibody-bound radioactivity as glucagon-1131equilibrates with, and is exchanged for, the moreplentiful nonradioactive glucagon. Evidence ofreversibility and an example of the rate at whichdissociation takes place are given in Figure 3.

It has been shown that the insulin-binding re-action obeys the law of mass action (8). At aconstant antibody concentration, the amount ofinsulin-binding increases with the quantity of in-sulin incubated, although the fraction of the incu-bated insulin that is bound declines. The resultsof the experiment shown in Figure 4 indicate thatglucagon-binding is similar in this respect.

2.4-

2. 2

° 2.0-Z .6-

z 14

0o 2-Z .

'J 1.0-0 so-S O

E .60-

R .40-

.20-

x

I~~~~~

KI.

-60

I"

-50 nz

-40

C)-30 0

z

0;-20 °

z

-10 X

0 2 3 6 a 10 2 14 16Concentrotion of Glucagon-l"' (,ugm / ml )

FIG. 4. THE EFFECT OF VARYING THE CONCENTRATIONOF GLUCAGON-I1" UPON THE AMOUNT (LEFT) AND PERCENT (RIGHT) OF GLUCAGON-I1.. BINDING TO ANTIBODY INUNDILUTED RABBIT ANTIGLUCAGONSERUM. The amountbound (0) increases nonlinearly, suggesting a "mass ac-tion" effect, while the percentage bound (X) declinesprogressively.

1284

GLUCAGONANTIBODIES AND AN IMMUNOASSAYFORGLUCAGON128

1J. InII m7l ln0assal'1:100 RABBIT ANTI-GLUCAGONSERUM

100 JJIGMS GLUCAGONI15157 8 % BOUNDTO ANTIBODIES

A

I

0:

2,000

PREINCUBATED

6.000 10.000

Standard curves. When unlabeled glucagon in5 per cent albumin solution is incubated withantiglucagon serum prior to incubation of theantiserum with glucagon-I'1 , the B/F ratio ofthe latter will fall. The magnitude of this fall isrelated to the quantity of the unlabeled glucagonthat has been added. When the B/F ratio isplotted as a function of the quantity of added un-labeled glucagon, a characteristic curve is obtainedwhich serves as the standard curve for the assay(Figure 5A, B). In general, an unequivocal de-cline in B/F ratio from the values obtained witha 5 per cent albumin blank is encountered when-ever the quantity of unlabeled glucagon approachesor exceeds the amount of glucagon-1131 in thesystem. Initially, relatively small amounts ofunlabeled glucagon cause a proportionately largefall in B/F ratio, but with larger amounts a pro-gressively smaller decline is noted and the curveapproaches the horizontal.

The ability of other peptide hormones, glu-cagon-free insulin, vasopressin, and ACTH, tolower the B/F ratio of glucagon-I131 was testedand compared with that of glucagon (Table II).Only glucagon caused a significant decline in theB/F ratio of glucagon-I3 , suggesting a high de-gree of specificity for the assay method.

Because it wvas not possible by means of cur-rent techniques and equipment to count less than50 to 100 /qg of glucagon-I'31, maximal sensitivitypossible under these circumstances was 50 to 100

untg of glucagon. In achieving this level of sensi-tivity, a 1: 100 dilution of the pooled antiserumwas employed (Figure 5B).

Although curves made with different lots ofglucagon-I13' and different lots of dilute anti-

Added Unlabeled Glucagon (,ppgm/ml)

FIG. 5. STANDARDGLUCAGONASSAY CURVE. A. Pre-incubation in a 1: 100 dilution of serum of increasing

amounts of nonradioactive glucagon, from 2,000 to 10,-000 ug causes a progressive decline in the B/F ratio of100 u~ug of glucagon-I31. A plot of B/F as a function ofthe amount of preincubated "cold" glucagon yields theabove characteristic curve which serves as a standardcurve in the glucagon assay. The high degree of repro-

ducibility is indicated by the proximity of the duplicatedeterminations.

B. Maximal sensitivity attainable by currently em-

ployed techniques is 50 uug. In the above curve, 50jugg of "cold" glucagon caused a significant fall in theB/F ratio of glucagon-I'31.

1.50.

1.35

1. 20-

1.05

.90-

B/F

.75 -

.60 -

.45-

.30

.15 -

0-

B

Ratio

I no wvv svI vw wov vv. v rrun.1 .", .JCOLD GLUCAGON( MIJGM/MwIL. I

1285

b

R. H. UNGER, A. M. EISE:NTRAUT, M. S. McCALL AND L. L. MADISON

TABLE II

Effect of various polypeptide hormones on B/Fratio* of glucagon-I13'

Glucagon-freeHormone insulin Glucagonconcen- Vaso-tration Exp. I Exp. II pressin ACTH Exp. I Exp. II

JI Ag/mI0 1.030 1.173 1.173 1.379 1.030 1.173

100 1.379 0.930200 1.424 0.796400 1.039 1.020600 0.753800 1.039

1,000 1.110 0.990 1.063 0.7744,000 0.2585,000 0.979 1.002 0.2826,000 0.202

10,000 0.956 0.988 1.143 0.129 0.15215,000 0.981 0.943 0.11420,000 0.983 1.154 0.075

* Ratio of antibody-bound to free glucagon-I'31.

serum differed from each other to a varying de-gree, excellent reproducibility was noted when thesame standards and other materials were setup in duplicate on the same day (Figure 5A).Under these circumstances, the standard devia-tion from the mean of 35 consecutive duplicateB/F ratios of glucagon standard solutions was

- 0.0436.In vivo recovery of exogenous glucagon. The

disappearance rate of beef-pork glucagon-I31 fol-lowing its intravenous injection in man has beenstudied previously by measurements of the levelof trichloroacetic acid-precipitable radioactivityin plasma. Measurements in this laboratory innormal humans reveal a rapid logarithmic declinein glucagon-I131 plasma concentration with a half-life of approximately 10 minutes. In order totest the ability of the immunoassay to measureexogenous beef-pork glucagon in vivo, 3 mg ofunlabeled glucagon was injected intravenously,and blood specimens were drawn at intervals for1 hour thereafter. Measurements of glucagonconcentration by means of the immunoassay gavea disappearance curve for exogenous glucagonquite similar in form to that obtained by radio-activity measurements after radioglucagon ad-ministration (Figure 6). The higher plateau oftrichloroacetic acid-precipitable radioactivity is at-tributed to the altered glucagon-I131 which bindsto serum proteins and disappears less rapidly thanundamaged glucagon-IJ1 .

Endogenous glucagon. Since an immunoassayfor glucagon would be employed primarily forthe study of endogenous glucagon levels in dogs

and in humans, it seemed important to determinewhether cross reactivity between rabbit antibodiesto beef-pork glucagon and canine or human glu-cagon exists. The pancreases of an anesthetizeddog and of three deceased humans were obtainedand immediately frozen. These organs were ex-tracted by the method of Sutherland and de Duve(17) and the dried residue reconstituted in chlo-ride-phosphate buffer containing 2 per cent albu-

100-

aCDv

- 180

-120 g

0

CD

OaC)3

n 30

neX

D

-60

0 "

3mg. I.V.60

Minutes

FIG. 6. GLUCAGONTIME-DISAPPEARANCE CURVE. Theplasma disappearance rate of 3 mg of glucagon-I1'" wasdetermined after its intravenous injection into a normalsubject (solid line) by measuring the trichloroaceticacid-precipitable radioactivity in plasma at multiple in-tervals for 1 hour after injection. In a second normalsubject 3 mg of nonradioactive glucagon was injectedand its rate of disappearance measured by means of theradio-immunoassay (dotted line). Although the absoluteconcentration of glucagon in mtg/ml differed widely ineach subject at any given time, the form of the two disap-pearance curves was quite similar when, as was doneabove, glucagon concentration was calculated as per centof the peak 2.5 minute concentration. The higher valuesfor glucagon-I.3. obtained after 10 minutes are the resultof persistent circulation of damaged glucagon-I131 whichdoes not disappear at a rate comparable with the un-damaged material, remaining bound to serum proteins.The blood glucose concentration of the patient who re-ceived nonradioactive glucagon is charted above; theother patient, whose blood glucose response curve is notcharted, had a more striking hyperglycemic response.

1286

GLUCAGONANTIBODIES AND AN IMMUNOASSAYFOR GLUCAGON

TABLE III

Measurements of endogenous glucagon in pancreatic extracts

B/F ratios and glucagon concentrations in ,sequivalents per ml

Human Human Human pancreas W.M.t Human Humanpancreas pancreas kidney liver

Dilution of extract Dog pancreas M.H.* R.M.* Head Tail R.M.* R.M.*

1:10 0.005 B/F 0.007 B/F 0.066 B/F 0.008 B/F 1.01 B/F 1.271 B/F>10,000 Pqug/ml >10,000 ,uig/ml >10,000,uqug/ml >10,000 tt/Ag/ml 0 Iujg/ml 0 /A/Ag/ml

1:20 0.310-

1:100 0.061 B/F 0.059 B/F 0.864 B/F 0.314 B/F6,800 ,utg/ml 3,150luOug/ml 620 ;45g/ml 1,880,qug/ml

1:400 0.391 B/F 1.149 B/F 0.817 B/F1.400 1,650 ,uug/ml 0,ruug/ml 710 ,uug/ml

1:800 0.890 B/F 1.244 B/F 1.209 B/F590 ,uug/ml 0,u4ug/ml 0,uMug/mlTotal glucagon content 22.4 ,ug 26-36.3 ,ug 3.7 jAg 8.8-13.4 jg 0 ,ug 0 Mgof organ (juEq)

* Obtained 2 hours post mortem.t Obtained 5 hours post mortem.

min to 1 ml per g of original wet weight. Ten to800-fold dilutions were then made with the samealbumin-buffer solution, and 1-ml samples ofthese dilutions were assayed for glucagon. TableIII reveals the profound depression of B/F ratiocaused by extracts of both human and canine pan-creas as compared with control organs. The esti-mated glucagon concentrations are expressed inmicromicroequivalents of beef-pork glucagon permilliliter. Calculations of the total glucagoncontent of pancreas of both species range from12.5 to 36.3 ,uEq of beef-pork glucagon.

The results clearly demonstrate the existence ofcanine and human glucagon and indicate that somecross reactivity with rabbit antibodies to beef-porkglucagon is present.

DISCUSSION

The data confirm the previously reported (9)antigenicity of beef-pork glucagon in rabbits.Furthermore, they indicate that in virtually everycategory examined glucagon-binding antibodiesappear to resemble insulin-binding antibodies.

Like insulin antibodies, glucagon antibodies areof the nonprecipitating type and can, therefore, bedetected only by such methods as the paper chro-mato-electrophoretic technique of Berson and co-workers (8), or the Skom-Talmage technique(15), each of which permits separation of free

glucagon-I13' from antibody-bound glucagon-I13 .Whereas the insulin-binding antibody of humansand guinea pigs is in the inter-yfl-zone, the glu-cagon-binding antibody of rabbits appears to mi-grate electrophoretically as a y-globulin. Kineticsimilarities between insulin-binding and glucagon-binding reactions were observed. First, the reac-tion of glucagon and its antibody was found to bereversible. Second, like insulin-binding, glucagon-binding was governed by the law of mass action;at a constant antibody concentration, the amountof glucagon-I131 bound increased as the concentra-tion of glucagon-I131, although the percentage ofglucagon-I131 bound decreased progressively.

The only previous suggestion of glucagon anti-genicity consists of a report by Carleton, Grebenand Schulman (21) of a hypotensive reaction fol-lowing an intravenous injection of glucagon in apatient who had previously had insulin shocktherapy. Since commercial insulin is contami-nated with glucagon, it is possible that this was aninstance of glucagon sensitivity in humans. Ex-cept for this unproven case, the present studiesconstitute the only evidence for glucagon anti-genicity in any species. Preliminary studies inour laboratory indicate the presence of smallamounts of nonprecipitating antibodies to glu-cagon in some insulin-treated diabetic patients(22). Although the clinical significance of this

1287

R. H. UNGER, A. M. EISENTRAUT, M. S. McCALL AND L. L. MADISON

evidence of glucagon antigenicity in man is doubt-ful, it may be well to keep this possibility in mindin puzzling cases of allergy ascribed to insulin.

The competitive inhibition by nonradioactiveglucagon upon the reaction between glucagon-J131and its antibody, fully predictable from the fore-going kinetic studies, provides the basis of theimmunoassay. If, in a system containing tracesof glucagon-J131 and dilute antiserum, increasingamounts of nonradioactive glucagon are added,the B/F ratio of radioglucagon will progressivelydecline. A plot of the B/F ratio as a function ofthe amount of nonradioactive glucagon added givesa characteristic curve which serves as the stand-ard curve of the assay.

In contrast to bioassay methods which measureonly "glucagon-like" activity, the radio-immuno-assay appears to be highly specific and exquisitelysensitive. The assumption of specificity is basedupon the demonstrated lack of cross reactivity be-tween glucagon antibodies and polypeptide hor-mones other than glucagon, i.e., insulin, vasopres-sin, and ACTH. The maximal sensitivity of thetechnique, at present 50 /Attg per ml, is greaterthan the maximal sensitivity of the earlier bioas-say methods. Furthermore, additional increasein sensitivity is presumably possible.

The marked competitive action in the assaysystem exerted by pancreatic extracts from bothdog and man not only provides specific evidenceof the presence of glucagon in both these speciesbut gives assurance of at least some cross reac-tivity between rabbit antibodies to beef-pork glu-cagon and both human and canine glucagon. Al-though the low values for total pancreatic glucagoncontent may suggest that cross reactivity is ratherlimited, they may reflect postmortem glucagondestruction or inefficiency of the extraction tech-nique. In any case, since some cross reactivityis a prerequisite to the measurement of endogenousglucagon in these species, it seems possible thatthe immunoassay for glucagon will permit the ex-ploration of the hormonal status of endogenousglucagon in health and disease.

SUMMARY

Monthly injections in rabl)its of l)eef-lpork glu-cagon in Freund's adjuvant resulted in the ap-pearance of nonprecipitating glucagon-binding

antibodies detectable both by the method of Bersonand Yalow and that of Skom and Talmage. As inthe case of insulin antibodies, the reaction betweenglucagon-I13' and glucagon antibodies was a re-versible one, governed by the law of mass action.

The preincubation of nonradioactive glucagonwith antiglucagon serum reduced binding of glu-cagon-I131 to a degree related to the quantity pre-incubated, thereby providing the basis of a radio-immunoassay for glucagon. Other peptide hor-mones, including insulin, did not influence glu-cagon-I131-binding, implying a high degree ofspecificity for the assay, which was found capableof measuring as little as 50 ,upug per ml of glucagon.

The assay was used to measure the disappear-ance of glucagon following its intravenous injec-tion in a normal man. The resulting curve re-sembled in form the plasma disappearance curveof glucagon-I'3 .

High glucagon concentrations, measurable inextracts of human and canine pancreas, indicatecross reaction between the glucagon of these spe-cies and antibodies to beef-pork glucagon, whichshould make possible the first specific explorationsof circulating endogenous glucagon in health anddisease.

ACKNOWLEDGMENTS

The authors wish to thank Drs. Solomon A. Bersonand Rosalyn S. Yalow for valued advice. They are in-debted to Mrs. Genevieve Thompson for technical con-tributions, to Mrs. Patsy Rupard and Miss Pam Youngfor stenographic help, to the late Mr. John Ash for design-ing the chromatography boxes, and to Mr. Walter F. Sul-livan and Clyde Tilton for the illustrations.

REFERENCES

1. Staub, A., Sinn, L., and Behrens, 0. K. Purificationand crystallization of glucagon. J. biol. Chem.1955, 214, 619.

2. Bornstein, J., Reid, E., and Young, F. G. The hy-perglycaemic action of blood from animals treatedwith growth hormone. Nature (Lond.) 1951, 168,903.

3. Staub, A., Springs, V., and Elrick, H. Effect of glu-cagon on renal excretion of electrolytes. Fed.Proc. 1956, 15, 361.

4. Tyberghein, J. M., and Williams, R. H. Assay forglucagon in rabbit p)lasma. Metabolism 1958, 7,635.

5. Makmaii, M. H., Makman, R. S., and Sutherland,E. W. Presence of a glucagon-like material in

1288

GLUCAGONANTIBODIES AND AN IMMUNOASSAYFOR GLUCAGON

blood of man and dog. J. biol. Chem. 1958, 233,894.

6. Berthet, J. Some aspects of the glucagon problem.Amer. J. Med. 1959, 26, 703.

7. Arquilla, E. R., and Stavitsky, A. B. Evidence forthe insulin-directed specificity of rabbit anti-insu-lin serum. J. clin. Invest. 1956, 35, 467.

8. Berson, S. A., Yalow, R. S., Bauman, A., Roth-schild, M. A., and Newerly, K. Insulin-I13. me-tabolism in human subjects. Demonstration of in-sulin binding globulin in the circulation of insu-lin treated subjects. J. clin. Invest. 1956, 35, 170.

9. Unger, R. H., Eisentraut, A. M., McCall, M. S.,Keller, S., Lanz, H. C., and Madison, L. L. Glu-cagon antibodies and their use for immunoassayfor glucagon. Proc. Soc. exp. Biol. (N. Y.) 1959,102, 621.

10. Berson, S. A. and Yalow, R. S. Isotopic tracers inthe study of diabetes. Advanc. biol med. Phys.1958, 6, 349.

11. Yalow, R. S., and Berson, S. A. Immunoassay ofendogenous plasma insulin in man. J. clin. Invest.1960, 39, 1157.

12. Pressman, D., and Eisen, H. N. Zone of localiza-tion of antibodies; attempt to saturate antibody-binding sites in mouse kidney. J. Immunol. 1950,64, 273.

13. McCall, M. S., Timm, L., Eisentraut, A. M., andUnger, R. H. The preparation of high specificactivity glucagon-I"3'. To be published.

14. Berson, S. A., Yalow, R. S., and Volk, B. W. Invivo and in vitro metabolism of insulin-I1" andglucagon-I.3. in normal and cortisone-treated rab-bits. J. Lab. clin. Med. 1957, 49, 331.

15. Skom, J. H., and Talmage, D. W. Nonprecipi-tating insulin antibodies. J. clin. Invest. 1958,37, 783.

16. Grodsky, G. M., and Forsham, P. H. An immuno-chemical assay of total extractable insulin in man.J. clin. Invest. 1960, 39, 1070.

17. Sutherland, E. W., and de Duve, C. Origin anddistribution of the hyperglycemic-glycogenolyticfactor of the pancreas. J. biol. Chem. 1948, 175,663.

18. Yalow, R. S., and Berson, S. A. Apparent inhibitionof liver insulinase activity by serum and serumfractions containing insulin-binding antibody. J.clin. Invest. 1957, 36, 648.

19. Mirsky, I. A., Perisutti, G., and Davis, N. C. Thedestruction of glucagon, adrenocorticotropin andsomatotropin by human blood plasma. J. clin. In-vest. 1959, 38, 14.

20. Berson, S. A., and Yalow, R. S. Quantitative as-pects of the reaction between insulin and insulin-binding antibody. J. clin. Invest. 1959, 38, 1996.

21. Carleton, J. L., Greben, S. E., and Schulman, J. L.Hypotensive reaction following the use of glucagon.A. M. A. Arch. intern. Med. 1957, 99, 817.

22. Unger, R. H., Sims, K., and Eisentraut, A. M.Unpublished observations.

1289